The European Commission, the European Union's executive arm, has officially confirmed it was hit by a cyberattack that misused AI. About a month after the attack was first detected, some internal systems have yet to be restored.
Bloomberg and the Associated Press reported on Nov. 19, local time, that the Commission said AI tools were used in last month's cyberattack and it is investigating the cause and scope of the damage. It is the first time the EU has officially confirmed it was a victim of an AI-enabled cyberattack.
Commission spokesperson Thomas Regnier (토마스 레그니에) said at a briefing that abnormal signs were first spotted on Oct. 22 and an initial investigation on Oct. 26 confirmed the attack. The Commission temporarily suspended several internal applications to prevent the damage from spreading. Most have been restored, but some services have yet to return to normal.
Regnier confirmed it was an AI-enabled cyberattack but did not disclose who was behind it, the purpose, or the exact scale of the damage. He said no indications of a data leak affecting all staff have been found so far.
Earlier, Politico reported the attack affected internal systems including payroll and working hours management, medical expense reimbursement and translation for at least 10 days. The Commission did not officially confirm which systems were affected or how long the disruption lasted.
The attack may have used ARTEX, an AI-based penetration-testing tool. CERT-EU, which handles cybersecurity for EU institutions, warned on Nov. 5 that ARTEX is being misused in real attacks. ARTEX is an open-source tool that uses AI to identify vulnerabilities and misconfigured security settings and automate the intrusion process.
CERT-EU assessed it is highly likely ARTEX was also used in similar attacks targeting institutions under its jurisdiction. The Commission did not officially name the tool used in this case.
The U.S. Federal Bureau of Investigation also issued a security alert on ARTEX on Nov. 13. In some attacks, it was found that it took less than 3 hours from initial target reconnaissance to gaining administrator privileges.
As AI-based attack tools automate everything from finding security weaknesses to intrusion, the cybersecurity burden on public agencies is also growing. The Commission plans to work with CERT-EU to check for further damage and continue security measures.